Ah fantastic, I was planning to do this at some point. Please fix rebase conflict (D4898 landed) but otherwise looks good, also include make bench-bitcoin (or whatever) in the test plan just to make sure that compiles too.

rebase

rebase onto D4906 for CI

rebase

rebase

note: this is missing IsSolvable changes (function did not exist at this time)

(removed again in D3136)

I agree with that rationale, in short "if you have any trusted copy of ABC, it contains the keys to verify any new version".

link to PR please

smaller comment

Hmm weird I just noticed a bunch of these comments now (the first block) ... is it possible they didn't show up before? Anyway... back to work...

Update for changes in ancestor, add note that this functions
equally well from two possible parents (one or the other).

Alternative: D4905 and D4906

drop totalsigops (unused)

update ; split off sigops mempool test additions to D4904

rebase, SigOps->SigOpCount

In D4902#118339, @deadalnix wrote:

In general, you'd want to patch to solve a problem. It is unclear here what that problem is.

In D4871#118376, @Mengerian wrote:

I'm not a fan of the new name... "Endian" to me seems hardware specific, and usually implies some power-of-2 number of bytes....

prefer name like 'sig op count'; caught a few more mentions of 'cost'

rm unused packageVirtualSize

rebase & update tests

get rid of sigopscost concept (new parent)

also rename virtualtxsize args

It looks like this depends on D4859 so make sure to note that in the summary as "Depends on D4859" -- phabricator should add a child-parent relationship.

comments

rebase for modified parent (no actual change)

copy to outparam rather than reset-modifying outparam

One possibility that comes to mind is:

• Calculate the coin-age of inputs as the lower of: BCH amount * number of confirmations, or BCH amount * N where N is some number in the 100s to 1000s. In other words, calculate coin-age as usual but clamp the age to some number ~ day-week length.
• When limiting mempool entry (when comparing to minfeerate), calculate the fee as: actual_fee + X * input_coin_age for some very small X.

move consensus-affecting bits into D4835 so this is only the flag effect

absorb things from D4617 and flip dependency

rebase for parent landing ; move decl down

sub 0 -> 0x00 to get linter to remove weird spaces

I also have the PR10252 and PR10488 as distinct commits and can make diffs from them (stacking off D4813) if that is preferred.

alt: D4829

Yeah fair enough, I was half-expecting this review. Actually the original PR 9602 did fully remove the argument, and I had no idea they restored the parameter as a dummy argument. I just invented the intermediate dummy step myself (the parent Diff) since it was a no-brainer.

tweak notes per comment

Simplify by removing the scriptlen/23 piece.
I've been racking my brain trying to think of any examples that it would
impact and I can't find any.